Function Calling Recommendation: Gemini 2.0 Flash Thinking
Function Calling → Gemini 2.0 Flash Thinking
Ikhtisar Kasus Penggunaan
Gemini 2.0 Flash Thinking by Google is a recommended option for the Function Calling use case.
Model Rekomendasi
| Peringkat | Vendor | Jendela Konteks | Score |
|---|
| #7 | Google | 1048K | 37/100 |
Analisis Kinerja
Kinerja Benchmark
| Benchmark | Skor |
|---|
| MMLU | 86.5 |
| HUMANEVAL | 87.2 |
| GSM8K | 91.3 |
| MATH | 56.6 |
| BBH | 87.9 |
| GPQA | 61.0 |
Kelebihan
- MMLU score 86.5, strong knowledge reasoning.
- HumanEval 87.2, excellent code generation.
- GSM8K 91.3, robust math reasoning.
- 支持文本、图像、音频多模态输入。
Persyaratan
- 需要 google 的 API 密钥
- 输入长度须在 1048K 上下文窗口内